Installation

Using pip

pip install scparadise -U

Create environment for using scparadise

conda create -n scparadise python=3.12

conda activate scparadise

pip install scparadise -U

If you want to use scParadise from R, you need to configure a Python environment in RStudio: Tools - Global Options - Python

Create environment from scparadise.yml (recommended)

1. Download scparadise_py3.12.yml or scparadise_py3.11.yml or scparadise_py3.10.yml.

2. Install g++ (optional, for a clean installation):

sudo apt update
sudo apt-get install g++

3. Execute the following command in Anaconda (from directory with scparadise_py3.12.yml or scparadise_py3.11.yml or scparadise_py3.10.yml):

conda env create -f scparadise_py3.12.yml

The installed environment is based on Python 3.12 (scparadise_py3.12) or 3.11 (scparadise_py3.11) or 3.10 (scparadise_py3.10) and includes the latest version of scparadise, scvi-tools, scanpy, muon, harmony, jupyterlab, liana, decoupler and other packages for scRNA-seq analysis.

GPU support

scParadise automatically uses an NVIDIA GPU if available. If you do not have a GPU, scParadise will fall back to the CPU. We recommend using GPUs to train custom models, as training on CPUs can take a long time.